Predicting An Outcome Associated With A Driver Of A vehicle

Abstract

Methods and systems are disclosed for predicting an outcome associated with a driver of a vehicle using a machine learning statistical model. The disclosed techniques include obtaining a plurality of input vectors for plurality of points in time, wherein each input vector includes a plurality of variables with a weight vector. Each variable represents data captured from a sensor or a data source. A training dataset for the machine learning model is created by capturing the values of outcome of interest for various values of each input vector for each point in time. The outcome of interest is the predicted by utilizing the machine learning model. In various embodiments, the predicted outcome of interest may be a risk or an energy consumption level associated with the driver.

What is claimed is: 
     
         1 . A system for predicting a value of an outcome associated with a driver of a vehicle, said system comprising:
 (a) an acquisition module that obtains a plurality of input vectors at defined time intervals at a plurality of points in time;   (b) at least one sensor that captures data associated with each input vector of said plurality of input vectors at each point in time for said vehicle, wherein each of said plurality of input vectors comprises a plurality of sensor data and a plurality of database data;   (c) a processor that predicts said value using a statistical model, wherein:
 (i) said value comprises a function of corresponding input vectors and an associated weight vector, 
 (ii) said weight vector is derived using said plurality of input vectors and associated values of said outcome at each point in time of said plurality of points in time, and represents an overall effect of each said input vector on said value, 
 (iii) said value is predicted through a regression analysis of said value associated with said each input vector, and 
   d) a user interface that displays results corresponding to said value of said outcome.   
     
     
         2 . The system of  claim 1 , wherein said at least one sensor is one of a microphone and a camera. 
     
     
         3 . The system of  claim 1 , wherein said sensor data contains driver behavior data. 
     
     
         4 . The system of  claim 3 , wherein a training of said statistical model utilizes cross-validation. 
     
     
         5 . The system of  claim 3 , wherein said value of said outcome represents a risk associated with said driver of said vehicle. 
     
     
         6 . The system of  claim 3 , wherein a risk scorecard for said driver of said vehicle is populated based on said value of said outcome. 
     
     
         7 . The system of  claim 6 , wherein an insurance rate for said driver of said vehicle is adjusted based on said risk scorecard. 
     
     
         8 . The system of  claim 3 , wherein a cruise control for said driver of said vehicle is regulated based on said risk level. 
     
     
         9 . The system of  claim 3 , wherein a location-based risk map for said driver of said vehicle is created based on said risk level. 
     
     
         10 . The system of  claim 9  wherein, one or more autopilot parameters for said driver of said vehicle are regulated based on said location-based risk map. 
     
     
         11 . A method for predicting a value of an outcome associated with a driver of a vehicle, said method comprising the steps of:
 (a) obtaining a plurality of input vectors for said vehicle at defined time intervals at a plurality of points in time, each input vector of said plurality of input vectors associated with each point in time of said plurality of points in time;   (b) capturing said value of said outcome corresponding to said each input vector at said each point in time for said driver of said vehicle, said each input vector comprising a plurality of sensor data and a plurality of database data;   (c) performing said predicting by using a processor and a statistical model, wherein:
 (i) said value comprises a function of corresponding input vectors and an associated weight vector, 
 (ii) said weight vector is derived using said plurality of input vectors and associated values of said outcome at each point in time of said plurality of points in time, and represents an overall effect of each said input vector on said outcome, 
 (iii) said value is predicted through a regression analysis of said value of said outcome associated with each said input vector, and 
   d) presenting results on a user interface corresponding to said value of said outcome.   
     
     
         12 . The method of  claim 11  providing said at least one sensor to be one of a microphone and a camera. 
     
     
         13 . The method of  claim 11  providing said sensor data to include driver behavior data. 
     
     
         14 . The method of  claim 13  utilizing cross-validation in a training of said statistical model. 
     
     
         15 . The method of  claim 13 , wherein said value of said outcome represents a risk level associated with said driver of said vehicle. 
     
     
         16 . The method of  claim 13  populating a scorecard for said driver of said vehicle based on said value of said outcome. 
     
     
         17 . The method of  claim 16  adjusting an insurance rate for said driver of said vehicle based on said scorecard. 
     
     
         18 . The method of  claim 13  regulating a feature of said vehicle based on said risk level. 
     
     
         19 . The method of  claim 13  creating a location-based risk map for said driver of said vehicle based on said risk level. 
     
     
         20 . The method of  claim 13 , wherein said value of said outcome represents a fuel consumption level associated with said driver of said vehicle.